Able But Unused — An Introduction to Learned Non-Use and CIMT

This article is the last of three parts dealing with treatment design in "Rehabilitation Basics". The preceding "Why Therapy Is Designed That Way — An Introduction to Motor Learning Principles" covered how to arrange practice, and "How Much Is Enough — An Introduction to Dose-Response in Rehabilitation" covered how much to do. This part looks at a case where those two are tied together into a single treatment protocol. A protocol is a treatment plan that determines in advance what, for how long, and in what order to do things.

The primer article "Neuroplasticity — A Brain That Builds New Roads Beside the Broken Ones" ended leaving behind a scale. On one side is practice that gets daily actions done with the sound hand. As the primer defined it, compensation is getting the same thing done by using another body part or another method instead of the lost way. On the other side is therapy that makes the injured side actively used. This article deals with where the second side of that scale, the therapy that makes the injured arm used, came from.

That therapy was not deduced from some theory. It came from a formulation, that is, a framework of explanation, built to explain a single observation.

Why You Need to Know — What Can Be Done and What Is Actually Used

A monkey whose sensory nerves in one front leg have been surgically removed will not use that leg, even though there are plenty of motor nerves left to move it (PMID 17039223). This state, where the sensory nerves are severed so that sensations from that leg cannot reach the brain, is called deafferentation. On the other hand, the nerves carrying instructions for the muscles to move remained intact. The path to move existed, but the animal did not move.

Learned nonuse is a state where a person or animal stops using a limb even though the ability to move remains, created not by the injury itself but by learning after the injury (PMID 17039223). The reason the word "learned" is attached to the name is explained as follows. Immediately after the injury, attempts to do things with that arm repeatedly fail, and using the healthy arm succeeds. When this experience is repeated, not using the affected side hardens into a habit, and that habit remains even later when the ability to move returns.

When two things that seem like they should go together—such as the ability to move and the degree to which it is actually used—behave independently, it is called dissociation. The same dissociation occurs in humans following nerve damage (PMID 17039223).

Here emerges the backbone of this entire article. The effects of nerve damage can differ significantly between two measurements. One is requested performance, which is the motor ability demonstrated in a laboratory when asked to "try this movement." The other is spontaneous use, which is the degree to which that limb is used on one's own in real life when no one asks. Therefore, these two values must be evaluated separately (PMID 17039223). There are also tools that measure the two separately. In treatment research addressing learned nonuse, the methods used to measure the two areas separately have been evaluated as reliable and validly measuring what they intend to measure (PMID 17039223).

The same distinction appeared in earlier pieces of this series. "Why Hold Several Rulers Against One Person — An Introduction to the ICF Framework" distinguished between what can be done (capacity) and what is actually done (performance), and "How Do We Measure Recovery — An Introduction to Rehabilitation Outcome Measures" asked what we use to measure. Learned nonuse is the place where that question stands out most sharply. For the same arm, the laboratory and daily life can yield completely different answers.

The Big Picture — So There Are Two Targets for Therapy

The formulation of learned nonuse yields one prediction. When rehabilitation therapy is requested, it can produce different effects on motor performance versus how much the more-affected arm is actually used in daily life (PMID 17039223).

CI therapy (constraint-induced therapy) is a treatment bundle that restricts the use of the less-affected arm, provides intensive training with the more-affected arm, and uses behavioral techniques to transfer that use into daily life. The version used for the arm is commonly called CIMT (constraint-induced movement therapy). The restriction is a device that temporarily blocks the easy path of getting things done with the sound arm, forcing the injured arm to be used.

The overview that formulated learned nonuse divides this therapy into two targets. This is an overview by the research team that developed the therapy. CI therapy improves performance upon request while focusing attention on spontaneous use. The same overview goes a step further, noting that spontaneous use is precisely where this therapy produces by far its greatest effect (PMID 17039223). The authors attribute that result to the "transfer package." The transfer package is a bundle of behavioral techniques that carries improvements made in the therapy room over into daily life. The authors note that this bundle can also be used alongside other treatments to carry improvements from the therapy room into daily life (PMID 17039223).

Following this explanation, summarizing CIMT as "a treatment that ties up the sound arm" causes this structure to vanish entirely. The definition in a scoping review points in the same direction. CIMT is a neurorehabilitation technique aiming to induce neuroplastic changes and overcome learned nonuse by combining neurological and behavioral mechanisms (PMID 37842361). Here, neuroplastic changes are those brought about by the ability defined in the primer "Neuroplasticity" — the ability of the brain to reorganize itself by forming new neural connections.

The restriction is only one component of CIMT, with the remainder being intensive training and behavioral techniques to transfer use to daily life. Modified CIMT (mCIMT) is a variation that focuses on the sensorimotor function of the affected arm (PMID 37842361).

The Direction of the Effect — Compared to Usual Rehabilitation

In stroke patients with hemiparesis, CIMT significantly improved arm function and activities of daily living more than usual rehabilitation (PMID 39649717). Hemiparesis is a condition where one side of the body loses strength or cannot move at all, which is the state covered in the primers "What a Stroke Actually Is — Blocked, Burst, and the Clock" and "Paralysis and Spasticity — Where the Command Breaks Shapes What Follows".

The results pointed in three directions. CIMT improved upper extremity motor function more than usual care, reduced upper extremity motor impairment more, and increased the performance of activities of daily living (ADLs) more (PMID 39649717). The upper extremity is the arm from the shoulder to the hand. Motor impairment is the extent to which the movements the body can perform are reduced due to injury, while motor function is the degree to which one can accomplish tasks with that arm. Activities of daily living are self-care tasks that a person performs on their own every day, such as eating, dressing, bathing, using the restroom, and transferring. "Significantly" means that the difference is hard to write off as a coincidence, though we must separately weigh how large that difference feels to the patient.

These results came from a meta-analysis gathering 34 randomized controlled trials (PMID 39649717). A randomized controlled trial is a test where participants are divided as if by drawing lots, comparing a group that received the treatment with a group that did not or received a different treatment. A meta-analysis is an analysis that pools the results of multiple trials and recalculates them all at once. The number of trials used for each outcome was 21 for upper extremity motor function, 13 for upper extremity motor impairment, and 12 for ADL performance. The researchers searched MEDLINE, Embase, and the Cochrane Library. The risk that the results leaned to one side due to the trial's design (risk of bias) was checked using Cochrane RoB 1.0. The certainty of the evidence was assessed using GRADE (PMID 39649717). The certainty of the evidence is a grade assigned by examining how much we can trust those results based on study design and consistency of findings. However, the abstract does not state which grade was assigned to each outcome.

The authors recommend applying CIMT by adjusting it to the strength of the affected side—meaning the injured arm (PMID 39649717). This means CIMT is not a fixed prescription applied with the same intensity to everyone, but a package adjusted to that person's arm condition.

What Kind of Protocol It Is — The Time Limit Is Set as a Range

The dose covered in "How Much Is Enough" appears here as a specific number. The protocol ranked first in upper limb motor recovery was CIMT, which combines trunk restraint and restricts the less-affected arm for at least 4 hours and no more than 6 hours a day (PMID 37039604).

This ranking came from a network meta-analysis ranking the relative efficacy of various CIMT protocols. A network meta-analysis is an analysis that lines up multiple treatments in a single row by connecting treatments that have never directly faced each other through a common comparator. The researchers included 44 out of 1,150 papers, involving 1,779 participants. The quality of the studies was assessed using the revised Cochrane risk of bias tool, and the analysis was a frequentist random-effects network meta-analysis (PMID 37039604).

Trunk restraint fixes the torso against the backrest of a chair, preventing the patient from leaning the trunk forward to make up the distance when reaching out an arm. The example given in the primer "Neuroplasticity" as a movement pattern compensation is precisely this: reaching out to touch a cup by leaning the shoulder and torso. Trunk restraint blocks that compensatory path, forcing the arm to move directly.

The metrics where this protocol led the way were the FMA-UE and the ARAT (PMID 37039604). Viewed through the classification of "How Do We Measure Recovery", the FMA-UE (Fugl-Meyer Assessment of Upper Extremity) measures the body's motor impairment, while the ARAT measures whether a person can grasp and move objects when asked to do so in a laboratory setting. Both fall on the requested-performance side.

The same condition appeared again in ADL improvements. Combining trunk restraint and restricting the less-affected arm for 4 to 6 hours a day significantly improved the Motor Activity Log (MAL) Quality of Movement (QOM) and Amount of Use (AOU) scales (PMID 37039604). The MAL (Motor Activity Log) is an interview-style scale that asks the individual how often and how well they use that arm in daily life.

Two things can be read from these results. First, the protocol ranked first was defined by a range of 4 to 6 hours a day. This means it was the highest among the protocols compared, and it is not a direct comparison result showing that exceeding 6 hours makes things worse. The authors also note only that it is worth considering in clinical practice (PMID 37039604). This time is the condition of the protocols compared in the studies, not a prescription that determines the right restriction time for an individual.

Second, the MAL, an ADL outcome measure, is a tool that measures spontaneous use. This protocol produced good results in both laboratory scores and real-life use, which leads us back to the two measurements established in the first section of this article.

Scope — Beyond the Arm and Stroke

CIMT and mCIMT are organized as effective for motor function recovery across various conditions such as cerebral palsy and traumatic brain injury, as well as stroke. Significant improvements have been reported in motor function and quality of life (PMID 37842361). This synthesis came from a scoping review. A scoping review is an overview that broadly maps what and how much has been researched on a topic, so its nature differs from a meta-analysis, which pools and calculates effect sizes. The same review notes the need for long-term studies, comparative and combination treatment studies, and exploration of less-studied conditions like multiple sclerosis and brachial plexus injury (PMID 37842361). This means evidence is still accumulating for conditions outside of stroke.

The scope does not stay with the arm, either. An overview formalizing learned nonuse groups several treatments of the same family together. They are lower-extremity CI therapy following stroke, spinal cord injury, or femoral fracture; CI aphasia therapy; focal hand dystonia in musicians; and phantom limb pain. And it notes that CI therapy is not a treatment for stroke, but a treatment for learned nonuse in general (PMID 17039223). The authors view that learned nonuse manifests in various conditions as motor impairments larger than what can be explained by the damage alone (PMID 17039223). From this perspective, what CI therapy targets is not the damage itself, but the disuse layered on top of it.

Common Misconceptions

"The core of it is tying it down."

Constraint is only one component. The very definition of this treatment is a combination of neurological and behavioral mechanisms, and the goal is overcoming learned non-use (PMID 37842361). If learned non-use is the problem, the place the treatment must target is use in daily life, rather than laboratory scores (PMID 17039223).

"The longer the tying time, the better."

What ranked first in a network meta-analysis was a protocol defined as a 4-to-6-hour-per-day band (PMID 37039604). This is a ranking, not a result proving an upper limit. There is no direct comparison result showing that going past 6 hours makes things worse. At the same time, evidence supporting "the longer, the better" is also absent here. What was pointed to as the top tier was a band with time bounded above and below.

"If it works, you might as well go hard and early."

This counterexample was already covered in "When Does Recovery Happen — What Time Did, and What Treatment Did". VECTORS is a single-site randomized trial testing CIMT during acute inpatient rehabilitation. It compared usual upper limb therapy, CIMT matched to that same amount, and much higher-intensity CIMT. The high-dose CIT group improved significantly less at 90 days, and the dose-matched CIMT group showed no significant difference from the control group (PMID 19458319). In a subset of participants who underwent imaging, signs that lesions grew larger due to the activity were not seen (PMID 19458319).

The authors read this result as an inverse relationship where more dose actually led to less improvement. They wrote that motor training cannot be assumed to be more beneficial just because the dose is high, especially in the early stages post-stroke (PMID 19458319). The effects shown by the meta-analyses we looked at earlier and this result hold true together. That a treatment is effective and when and how much it should be done are separate questions.

To Go Deeper

• When Reading a Paper — The first thing to check is whether the outcome measure assesses requested performance (such as the FMA or ARAT) or spontaneous use (such as the MAL). If the treatment addresses learned non-use and an index of spontaneous use is missing, that absence itself is information.
• When Reading a Protocol — The things to look for are whether time limits and trunk restrictions were used together, the training duration, and whether behavioral techniques for transferring to daily life are written down. As we saw in "Why Therapy Is Designed That Way", such practice conditions are frequently omitted from papers.
• If You Have Read This Series This Far — You will read a single rehabilitation paper in this order: how recovery is possible → what it was measured with → whether the improvement is deep → was it time or treatment → how the practice was arranged → how much was done → and whether the person actually uses that arm in reality.

Source Papers

• The learned nonuse phenomenon: implications for rehabilitation. PMID 17039223 (2006)
• Pertinence of Constraint-Induced Movement Therapy in Neurological Rehabilitation: A Scoping Review. PMID 37842361 (2023)
• The Effect of Constraint-Induced Movement Therapy on Arm Function and Activities of Daily Living in Post-stroke Patients: A Systematic Review and Meta-Analysis. PMID 39649717 (2024)
• Effect of different constraint-induced movement therapy protocols on recovery of stroke survivors with upper extremity dysfunction: a systematic review and network meta-analysis. PMID 37039604 (2023)
• Very Early Constraint-Induced Movement during Stroke Rehabilitation (VECTORS): A single-center RCT. PMID 19458319 (2009)